1: \begin{abstract}

2: We show that smoothness implies \nci : the smoothness of an element

3: $a$  in

4: a Banach algebra with a one-parameter automorphism group is preserved

5: under inversion, and the norm of the inverse $a^{-1}$  is controlled by

6: the smoothness of $a$ and by spectral data.  In our context smooth

7: subalgebras are obtained with the classical constructions of

8: approximation theory and resemble spaces of differentiable functions,

9: Besov spaces or Bessel potential spaces. To treat ultra-smoothness, we

10: resort to Dales-Davie algebras. Furthermore, based on  Baskakov's

11: work, we derive explicit norm control estimates for infinite matrices

12: with polynomial off-diagonal decay. This is a 

13: quantitative version of Jaffard's theorem. 

14: 

15:  

16:  % For smoothness spaces we obtain iterated and generalized quotient rules that imply norm controlled inversion

17: % \begin{theo}

18: %   If \mA\ is a \BA\, and $\operatorname{smooth}_r(\mA)$ is a smoothness space (precise definitions follow) of smoothness $0 < r< 1$ then

19: % \[

20: % \norm{\inv A}_{\operatorname{smooth}_r(\mA)} \leq C_r \norm{\inv A}_\mA^2 \norm{A}_{\operatorname{smooth}_r(\mA)} 

21: % \]

22: % \end{theo}

23: % Here  $\operatorname{smooth}_r(\mA)$ can be one of the following:

24: % \begin{itemize}

25: % \item [-] a Besov space

26: % \item [-] a Bessel potential space

27: % \item [-] a space $C^k(\mA)$

28: % \item [-] a Dales-Davie algebra (this includes the previous item)

29: % \end{itemize}

30: \end{abstract}

31: